c.f this question on O() notation.

Stackoverflow actually has the 'homework' tag, and is willing to entertain questions of that kind. MO has strong prohibitions in place against it. The first problem here is that it's not entirely obvious that this is a homework question, and that changes my response.

  • If it's not a homework question, then I'm willing to nudge the OP along in comments to find the answer themselves
  • if it is, I'd rather not help.
  • either way, I'm loathe to actually POST an answer, because it's a gratuitous rep honeypot, and it might only encourage more questions along these lines, and I'm not sure I want to encourage questions that are so low-level.

But I recognize that there are differences of opinion on how welcoming we should be. So what say all ?

  • 5
    $\begingroup$ I think the question on O notation is simply far too basic, and that makes it clearly off-topic. Hence I've voted to close it. I think we can revisit the homework issue when we see a real on-topic question that looks like homework... $\endgroup$ – Jukka Suomela Sep 1 '10 at 19:18

I don't like the idea of answering homework questions because it's disingenuous (err...cheating?). The only challenge is determining when a question is homework.

I would similarly frown on questions where the OP asks: "here's a problem that I have; provide the solution and code". Questions should be based on knowledge sharing, not on shirking. It goes against the spirit of the site.

The question that sparked this discussion is in this vein:

"Please check and comment."

There is no evidence of engagement in the subject. The question itself isn't specific about anything other than reviewing someone else's work.

  • 11
    $\begingroup$ "Questions should be based on knowledge sharing, not on shirking". That needs to be in the FAQ - brilliantly put ! $\endgroup$ – Suresh Venkat Sep 1 '10 at 19:31

I think the example question above has no place here. The question doesn't really meet the standards layed out in the FAQs. Whether or not said question is homework is irrelevant. This is layed out pretty well in your third bullet. If we don't want to encourage people to post answers to questions like this, why have it on the stackexchange? Of course this opinion does leave open the question about homework questions in general, and how to deal with them.

I think that homework questions, if they are sufficiently tricky, may very well make good questions. I also don't think we need to follow the MathOverflow model so closely as to copy its policy on homework questions. I say, judge the question on its own merits. If it's an interesting question or if an interesting answer can be posted despite the question it should be allowed.

Since homework questions don't tend to be interesting (there certainly are some good exceptions, especially from a class with a good professor) this is a self-enforcing protocol.

  • 4
    $\begingroup$ +1. Homework or not is irrelevant for this question. It is too basic to be on this site. $\endgroup$ – Robin Kothari Sep 1 '10 at 19:28

Regarding homework questions, I think we should be skeptical of a question if any of us think it is homework. In complexity theory (for example) there are tons of folklore theorems which are common homework problems. I have even referred to some of these in my answers. But I stopped short of explaining those parts carefully, in the faint hopes that these problems can continue to be used in homeworks.

That said, I can also understand the argument that maybe we shouldn't care about whether homework questions are posed, so long as the question is interesting. If a student wants to cheat it is all too easy to do so, and nowadays courses are designed so that homework either plays a very small role in grades, or the questions are devilishly designed to defeat the next generation of search engines.

Still, I won't post an answer if I think it's homework, or if I know it has been homework somewhere in the past. And I'll point out when this is the case. I think the basic principle of sticking to mostly research-level questions (or, at the very least, research-level answers) is totally compatible with this.


Now we've started having more HW level questions. See

Using decision version of TSP to solve optimization version

Is it possible to have a 4-coloring for a non-planar graph ?

We need to have a policy for such questions.

I think we should stick with MO's policy and close such questions if they are basic. If they are not of the advanced undergrad / grad student level, questions should be closed.

  • 2
    $\begingroup$ I am ok either way, as long as there is a clear policy, but I was disturbed by the unnecessarily humiliating tone of a couple comments on the "non-planar-graph" question you linked to. "Be Nice" is explicitly in the FAQ, and if we decide to close homeworky questions, we should close them diplomatically. $\endgroup$ – Aaron Sterling Sep 8 '10 at 15:40
  • 1
    $\begingroup$ The problem is that my homework question is your 'question I came across while reading some advanced material'. Like in the tsp question, I like the idea of not answering, but gently encouraging the OP to either expand on their thinking, or nudge them in the right direction. $\endgroup$ – Suresh Venkat Sep 8 '10 at 16:23
  • 1
    $\begingroup$ I think these questions are off-topic simply because they are too basic; hence I have voted to close. $\endgroup$ – Jukka Suomela Sep 8 '10 at 16:56

I think we should not answer any question which is an exercise in a textbook. Even giving hints is not appropriate in my opinion.


IMHO, homework questions are not welcome, and I discourage others from answering homework questions, specially from new/low reputation users using nicknames. If OPs want a solution for such a problem, they can consult textbooks or ask their course instructor/TA.

Also note that in most universities, getting help in solving assignments and presenting work which is not yours can have heavy punishments.

ps: I agree that there are exceptions where a homework question is acceptable.

  • 1
    $\begingroup$ This is a bit extreme. In some cases, the more difficult exercises in textbooks are open research problems. $\endgroup$ – Jukka Suomela Sep 2 '10 at 8:43
  • 1
    $\begingroup$ Let's exclude those open problems. But designing good homework questions is already not easy, I don't like making it even more difficult. Some books started asking explicitly that the answers to the exercises should not be posted on the web. $\endgroup$ – Kaveh Sep 2 '10 at 13:17
  • $\begingroup$ I agree with @Kaveh; explicitly stating "problem number 5 on page 300 of blah text" shouldn't be allowed. To the extent that these kinds of homework questions are generalizable, the OP should make an effort to ask them of their own accord rather than presenting it as a means to do their work for them. $\endgroup$ – Shane Sep 2 '10 at 15:50
  • $\begingroup$ @Kaveh: While I agree that this is not the place for homework questions, there is certainly a clear line between seeking more information on a problem/discussing it with others and plagiarism. My students discussing a particular problem and coming to a unanimous conclusion that proof by contradiction is the way to tackle is is a much different situation to receiving work from 3 separate individuals with identical errors of sign throughout. $\endgroup$ – Joe Fitzsimons Oct 26 '10 at 14:37
  • $\begingroup$ There is one exercise in Alon and Spencer which I have been unsuccessfully trying to figure out for years. I think cases like this should be exceptions. $\endgroup$ – Peter Shor Oct 26 '10 at 16:48
  • $\begingroup$ @Joe Fitzsimons: I agree with you, but I think the decision about this should be made by the course instructor. Some instructors allow some amount of discussions, for example as long as each students writes his/her own understanding of the solution independent of others, others don't don't even permit this. I don't think we are in a position to make the decision about the appropriate amount of hint in place of another instructor, she/he might have a specific objective in mind for asking it.IMHO asking a homework question is more like asking for the solution from another prof in the university. $\endgroup$ – Kaveh Oct 26 '10 at 19:13
  • $\begingroup$ @Peter Shor: I have already said that I agree that there are exceptions. :) My answer is mainly about elementary/textbook questions posted by new/low-reputation users using nicknames, i.e. an exercise given in an undergraduate course. I don't think any instructor would give an exercise to undergraduates which can't be solve by you. :) $\endgroup$ – Kaveh Oct 26 '10 at 19:18
  • $\begingroup$ If somebody does ask about a homework question (for these exceptions) should we encourage them to leave out the reference (book page and title), or not? $\endgroup$ – Peter Shor Oct 26 '10 at 19:46
  • $\begingroup$ Kaveh, as stated somewhere else: We are not responsible for other people abusing sources of knowledge (if that makes even sense). You could also argue that somebody who get the idea for his current research problem reading an answer here and gives no credit does wrong. I would rather consider stuff that is written here "private communication". $\endgroup$ – Raphael Oct 26 '10 at 22:11
  • $\begingroup$ @Peter Shor: I am not sure, but I think in those exceptional cases having a reference is useful, unless the authors of the book ask explicitly against posting answers of the questions on the web, e.g. Arora and Barak's book. $\endgroup$ – Kaveh Oct 26 '10 at 22:30
  • $\begingroup$ @Raphael: This is your view, I don't agree with it. I have told you, I can give you pointers to policy used in some top CS departments. I am going to write a more complete answer to the things you have mentioned under the comments of those two closed posts as soon as I find some free time. But for now: $\endgroup$ – Kaveh Oct 26 '10 at 22:37
  • $\begingroup$ 1. Aiding another person in committing an academic offense or dishonesty is itself an academic offense, and can be punished if you are in the same institute, and if you are not it can be forwarded to whatever institute you are in, i.e. you are responsible if you deliberately help others commit an academic offense. $\endgroup$ – Kaveh Oct 26 '10 at 22:37
  • $\begingroup$ 2. As I said above, the situation is similar to undergrad students asking profs to solve or give hints for assignments given to them by another course instructor. I don't think that is considered OK, similarly asking homework questions on this site is not OK in my opinion. $\endgroup$ – Kaveh Oct 26 '10 at 22:41
  • $\begingroup$ Kaveh, as for 1.: "deliberately" is the key word. We cannot know for what purpose and in what context of rules somebody asks, so we can by definition not deliberately help him commiting an offense. $\endgroup$ – Raphael Oct 27 '10 at 7:43

I'm not against homework questions where the qner is upfront about the fact, shares what they understand and don't understand about the solution (both SO guidelines), and where the question is at an appropriate level.

In general, the magic shield against the bad kind of homework question is for answerers to ask themselves "Do I know why the questioner wants to know the answer?" If not, don't rush to answer, ask why. For real questions, that typically casts light on how to go about providing an answer. For abusive questions, it is hard to give good reasons. I want something along these lines to be in the FAQ, along with what Shane said about the spirit of the site.

In this respect, we could do with being better about questions. I note that one of my own questions is not well-motivated: I should go off and fix it.

  • $\begingroup$ You didn't need really to add motivation ... that question was sufficiently self-motivated to be acceptable (although, as you point out, if you had added the motivation, you might have gotten a more relevant answer). $\endgroup$ – Peter Shor Oct 26 '10 at 16:45

Why do we need to agree on this now? As long as the site is not flooded by annoying homework questions, there is no problem. We should first wait what questions appear in practice and then discuss about their appropriateness.


While undergrad homework is allegedly not within the scope of this site, you can have assignments in graduate courses that are definitely not basic. For example, I am giving a talk in the near future to researchers about something I worked with in my Bachelor's thesis. So it might be hard to draw the line.

My proposal: Let us create a tag 'homework' if it does not exist already. This way, homework questions that are well posed (ref this discussion) can be allowed while people not interested in them can easily filter them out via the form on the main questions page. Something like https://cstheory.stackexchange.com/questions/tagged/-homework should yield a static page (with corresponding RSS feed) but I did not figure it out exactly. Anyway, this could be implemented easily if we went for it.

Edit: In response to the comments, let me elaborate why I want basic TCS questions here, not on MO.

  1. Students looking for help in TCS naturally go to the TCS site, not MO.
  2. I would expect more TCS expertise here than on MO (if not now because of the recent split, then certainly later on)
  3. By refusing to answer basic questions -- often in a less than nice way -- we leave a bad impression.
  4. No pity for people that prefer rejecting questions over using existing tools for filtering. One should be able to expect regular users to use filters rather than newcomers to know the scopes of at least two different communities.
  • 6
    $\begingroup$ But there is a website that exactly serves that porpose, math.stackexchange.com. They have a "homework" tag, and welcome TCS questions. Besides, the consensus is that we want this to be a research-level questions website, and thus basic questions are disallowed. It's not a homework vs non-homework issue, it's an issue of how basic the questions are. Everyone understands that a homework question in an advanced course can be difficult. $\endgroup$ – Robin Kothari Oct 26 '10 at 13:42
  • 2
    $\begingroup$ I'm with Robin. While we might seem overly paranoid at times, the problem with homework questions is that they open the floodgates, and most users aren't adept enough to use filtering to get rid of them. And as a user, I don't want to see homework questions on this site. $\endgroup$ – Suresh Venkat Oct 26 '10 at 15:21
  • $\begingroup$ Please see my edited answer. Suresh, we should be careful with extrapolating what a dozen power user wants to the whole community. We have over 1000 users right now. If we had 50 registrations every month by people who expect help for undergrad questions this would certainly change the site's scope by definition, right? $\endgroup$ – Raphael Oct 26 '10 at 16:34
  • 5
    $\begingroup$ actually no. I would leave this site if it was overrun by people wanting homework questions asked. And since I don't want to leave this site, I'll continue to vote against such questions :). this is not a democracy: it is entirely appropriate that active (power) users have more of a say, and the precise problem with homework questioners is that they are drive-by: they ask the question and leave, and are not part of the community. $\endgroup$ – Suresh Venkat Oct 26 '10 at 17:18
  • $\begingroup$ It is a scaled democracy. This implies that if a certain number of active users would be willing policies could change. Anyway, your assumption ("ask and leave") implies that we can as well answer and leave a good impression; writing "Pumping Lemma" as answer takes probable less time than close-voting and writing a comment. On the other hand, if we answer such questions the same user might ask another one or two homework questions but then become an active member. Nobody wants you to be a babysitter. Students can answer students, you know. This possibility is actively discouraged right now. $\endgroup$ – Raphael Oct 26 '10 at 17:38
  • $\begingroup$ On a related thought, we could introduce and encourage use of hardness specifying tags like 'basic', 'advanced', 'hard', 'open', each possibly combined with 'reference-request'. $\endgroup$ – Raphael Oct 26 '10 at 17:42
  • 2
    $\begingroup$ So why not just redirect people to math.SX then, instead of just closing the question? Or better yet, just migrate these questions once it is possible. That would be friendlier and everybody gets what they want. $\endgroup$ – Matthias Oct 26 '10 at 17:57
  • 3
    $\begingroup$ From Area51: "Stack Exchange sites are free to create and free to use. All we ask is that you have an enthusiastic, committed group of expert users who will check in regularly, asking and answering questions." We decided that experts would be more enthusiastic if they saw (for example) questions of the sort Peter Shor has been posting recently, than if they saw homework questions. MathOverflow has been very successful, and we decided to imitate most of their question-criteria. Also: meta.cstheory.stackexchange.com/questions/316/… $\endgroup$ – Aaron Sterling Oct 26 '10 at 18:45
  • 1
    $\begingroup$ Elementary questions do not need expert knowledge for answering, the corresponding course instructor or TA can answer them for sure, and let me also state that I don't buy the argument that "we cannot contact the instructor at the moment and we need an answer right now". $\endgroup$ – Kaveh Oct 27 '10 at 0:50
  • 3
    $\begingroup$ I also agree with the others, like MO this is an elitist site, it is for research-level question. I am also against homework questions (with some reasonable exceptions) for other reasons, but up to this point, as far as I remember, all homework questions we have closed were too elementary for the site, i.e. they were not research-level questions. If you are not happy with Math.SX and want another site for undergraduate students to ask elementary/homework questions, you are welcome to start one, but I am not going to spend my time there. $\endgroup$ – Kaveh Oct 27 '10 at 1:00

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .